基于OBE-ELM的球磨机料位软测量

发布时间：2018-04-30 05:20

本文选题：球磨机料位 + 软测量　；参考：《中北大学学报(自然科学版)》2017年05期

【摘要】：针对采用传统极限学习机在球磨机料位软测量建模过程中,存在鲁棒性差,预测精度不高等缺点,提出一种基于最优定界椭球(Optimal Bounding Ellipsoid,OBE)改进极限学习机(Extreme Learning Machine,ELM)的建模方法.该方法以球磨机振动信号为观测变量,采用偏最小二乘法提取有效特征,将提取到的有效特征输入到ELM中进行模型训练,并利用OBE在模型误差未知但有界的条件下,对网络权值进行约束优化.通过小型球磨机实验表明,在对球磨机料位进行回归预测时,该方法的评价指标与其它方法相比有所提高,测量结果的箱线图也直观展示该方法具有更好的鲁棒性.
[Abstract]:In view of the disadvantages of the traditional extreme learning machine in the soft sensor modeling of ball mill level, such as poor robustness and low prediction accuracy, a modeling method based on optimal bounded ellipsoid Bounding optimal Bounding Ellipsoid OBEBE-based improved extreme Learning Machine (ELM) is proposed. In this method, the vibration signal of ball mill is taken as the observation variable, the effective feature is extracted by partial least square method, the extracted effective feature is input into ELM for model training, and the model error is unknown but bounded by OBE. The network weights are optimized with constraints. The experimental results of small ball mill show that the evaluation index of this method is improved compared with other methods, and the box diagram of measurement results shows that the method has better robustness.
【作者单位】：太原理工大学信息工程学院;
【基金】：国家自然科学基金资助项目(61450011) 山西省煤基重点科技攻关项目(MD2014-07) 山西省自然科学基金资助项目(20150110052)
【分类号】：TH69

【相似文献】